The rate of recruitment to clinical trials is so low as to pose a major problem for the development of new cancer treatments. Moreover, serious ethical questions have been raised about a system which limits access to some of the most advanced treatments to a small proportion of cancer patients. Previous research seems to indicate that low accrual rates are primarily due to 1) the failure of many trial participating physicians to recommend or even discuss clinical trial participation with their eligible patients; and 2) generally low physician participation in rates in clinical trial groups. The proposed research seeks to describe physicians' treatment and referral decision-making for post-surgical breast cancer patients in the Pittsburgh SMSA as a model system. Approximately 600 patients will be selected from the case records of all consenting physicians who provide primary treatment to breast cancer patients in the Pittsburgh SMSA. Data will be gathered by a variety of techniques including intensive semi-structured interviews, medical record reviews and structured ratings for pre-selected decision-making factors. Data analysis will combine quantitative and qualitative content analysis of text. This content analysis will be greatly facilitated by the use of the UNIX Text Analyst (UTA), a computer program recently developed by one of the investigators on this project. Subsequent statistical analysis will initially rely on bivariate analyses to determine factors which are significantly correlated with the decision-making process. Discriminate analysis will be carried out to test how the selected description variables discriminate between physicians who do or do not participate in clinical trials. If we are to justify both the human and monetary expense of such clinical trials, the goal of all researchers should be to recruit subjects as widely and effectively as possible. Findings from this study will provide a focused knowledge based upon which other studies may test generalizability of findings to treatment of other tumors. These findings should also suggest ways to enhance trial participation by community physicians which can then be tested in future community intervention studies.